Based on the area of residence of the majority of the physician’s patients, practice location was identified and dichotomized into lower income area (comprised of the two lowest of the five neighbourhood income groups) and elsewhere. [...] Because the number of physician visits was skewed to the right, we employed the log transformation on the frequency of visits before proceeding with the multiple regression analysis. [...] Results were reported for this half of the patients (Table 3), and validated on the second half of the patients to confirm the stability of the estimates. [...] To examine whether the recall rate for hypertensive patients was indicative of the physician’s recall rate in general, the rate for hypertensive patients and the rate for other patients (excluding the hypertensive patients) was compared for each physician. [...] Neighborhood income and most of the health-related patient characteristics were assessed for their association with the number of physician visits and the chances of making at least ten physician contacts.